Evaluation of Machine Learning Approaches for Predicting Warfarin Discharge Dose in Cardiac Surgery Patients: Retrospective Algorithm Development and Validation Study

BACKGROUND

Warfarin dosing in cardiac surgery patients is complicated by a heightened sensitivity to the drug, predisposing patients to adverse events. Predictive algorithms are therefore needed to guide warfarin dosing in cardiac surgery patients.

OBJECTIVE

This study aimed to develop and validate an algorithm for predicting the warfarin dose needed to attain a therapeutic international normalized ratio (INR) at the time of discharge in cardiac surgery patients.

METHODS

We abstracted variables influencing warfarin dosage from the records of 1031 encounters initiating warfarin between April 1, 2011, and November 29, 2019, at St Michael's Hospital in Toronto, Ontario, Canada. We compared the performance of penalized linear regression, k-nearest neighbors, random forest regression, gradient boosting, multivariate adaptive regression splines, and an ensemble model combining the predictions of the 5 regression models. We developed and validated separate models for predicting the warfarin dose required for achieving a discharge INR of 2.0-3.0 in patients undergoing all forms of cardiac surgery except mechanical mitral valve replacement and a discharge INR of 2.5-3.5 in patients receiving a mechanical mitral valve replacement. For the former, we selected 80% of encounters (n=780) who had initiated warfarin during their hospital admission and had achieved a target INR of 2.0-3.0 at the time of discharge as the training cohort. Following 10-fold cross-validation, model accuracy was evaluated in a test cohort comprised solely of cardiac surgery patients. For patients requiring a target INR of 2.5-3.5 (n=165), we used leave-p-out cross-validation (p=3 observations) to estimate model performance. For each approach, we determined the mean absolute error (MAE) and the proportion of predictions within 20% of the true warfarin dose. We retrospectively evaluated the best-performing algorithm in clinical practice by comparing the proportion of cardiovascular surgery patients discharged with a therapeutic INR before (April 2011 and July 2019) and following (September 2021 and May 2, 2022) its implementation in routine care.

RESULTS

Random forest regression was the best-performing model for patients with a target INR of 2.0-3.0, an MAE of 1.13 mg, and 39.5% of predictions of falling within 20% of the actual therapeutic discharge dose. For patients with a target INR of 2.5-3.5, the ensemble model performed best, with an MAE of 1.11 mg and 43.6% of predictions being within 20% of the actual therapeutic discharge dose. The proportion of cardiovascular surgery patients discharged with a therapeutic INR before and following implementation of these algorithms in clinical practice was 47.5% (305/641) and 61.1% (11/18), respectively.

CONCLUSIONS

Machine learning algorithms based on routinely available clinical data can help guide initial warfarin dosing in cardiac surgery patients and optimize the postsurgical anticoagulation of these patients.

Advancing Practice-Based Research among Nursing and Health Disciplines Professionals by Creating a Network of Leaders

Representatives from various practice-based research programs have come together to establish a Canadian Practice-Based Research Network (CP-BRN). CP-BRN is a collective of healthcare leaders focused on identifying approaches and leveraging resources to support clinician-led research to advance evidence-based practice. This paper presents an overview of the development of the CP-BRN, the proceedings from the inaugural meeting of CPBRN members, and recommendations for nursing and allied health profession leaders considering establishing their own practice-based research programs. Next steps for the network are to raise awareness of its mission, expand the network membership as to grow its influence among healthcare leaders and to further advance evidence-based practice across both healthcare and academic institutions.

Effectiveness of a manuscript writing workshop on writing confidence amongst nursing and health disciplines clinicians

BACKGROUND

Dissemination of practice-based research findings is critical to advancing evidence and improving practice. While frontline clinicians are well-positioned to identify gaps in practice-based evidence, many barriers exist that challenge their ability to write and submit manuscripts for publication.

PURPOSE

Our study examined whether a manuscript writing workshop effectively increased nursing and health discipline clinicians' self-perceived confidence in manuscript writing.

METHOD

Participants recruited from an ongoing manuscript writing workshop completed an assessment tool at the beginning and end of each session. Thirty-one assessment tools were completed.

RESULTS

Participants reported higher levels of confidence following participation in the manuscript writing sessions. They also noted high levels of satisfaction with the session.

CONCLUSIONS

A manuscript writing workshop providing a supportive environment, mentorship, protected time, and quiet space is an effective way for leadership to increase confidence in manuscript writing amongst nursing and health disciplines clinicians.

Perceptions on implementation of a new standardized reporting tool to support structured morning rounds: Recommendations for interprofessional teams and healthcare leaders

Structured morning rounds have been used to improve communication, provide learning opportunities, and support patient care in various healthcare settings. The leadership team in an inner-city Inpatient Mental Health Unit identified a gap in the structure of morning rounds; to remedy this, a standardized reporting tool and structured morning rounding process were implemented. This short communication reports on an evaluation of staff...s perceptions on the outcomes of using a standardized reporting tool and structured rounding process to improve efficiency and communication regarding patient care among an interprofessional team. Feedback was provided on logistics, attendance, supporting factors that ensured ease of use for the new structure, and benefits of the program. While this evaluation focuses on the opinions of nurses, it serves as an example for leadership in various healthcare units on the benefits of a structured rounding program, and key factors that contribute to making the process successful. This short communication also provides an example for an efficient communication tool that can be adapted to meet the needs of various groups of healthcare disciplines.

The educational needs of Canadian homeless shelter workers related to traumatic brain injury

BACKGROUND

Traumatic brain injury (TBI) has a higher prevalence in the homeless population. Caregivers to individuals who have TBIs may require better education surrounding screening, diagnosis and management of this disease to tailor interventions to their clients' needs.

OBJECTIVE

To assess the insight and educational needs of homeless care providers in recognizing and dealing with clients who had experienced a TBI.

METHODS

A survey assessing the point of views of homeless care providers across Canada regarding their level of confidence in identifying and managing symptoms of TBI.

RESULTS

Eight-eight completed surveys were included. Overall, frontline workers expressed a moderate level of confidence in identifying and managing TBI, stating that educational initiatives in this context would be of high value to themselves and their clients.

CONCLUSIONS

Frontline workers to homeless clients rate their educational needs on the identification and management of TBI to be high such that educational initiatives for shelter workers across Canada may be beneficial to increase their knowledge in identifying and managing the TBI-related symptoms. Improved education would not only benefit frontline workers but may also have a positive effect on health outcomes for their clients.

2072. Do ID and Non-ID Clinicians Agree on IV to PO Switch Criteria? Results from Phase I of the INForming IV to ORal MEdication Report (INFORMER) Project

Background

Converting intravenous (IV) antibiotics to an oral (PO) route is an important stewardship activity to reduce patient harm, including extravasation, thrombophlebitis, and catheter-related infections. The INFORMER Project aims to develop a “smart” electronic tool to streamline IV to PO conversion in eligible patients using an algorithm derived from patient-level data. In designing the algorithm, we noted significant clinician subjectivity in reviewing PO eligibility criteria. To support algorithm development and frontline clinician buy in for future e-tool use, an initial step of our project explored agreement level for IV to PO switch between general internal medicine (GIM) vs. ID clinicians.

Methods

A convenience sample of GIM patients (tertiary teaching hospital) were reviewed in a 4-month pilot. Patients were still on the ward and received a target IV antibiotic (fluoroquinolone, sulfamethoxazole/trimethoprim, clindamycin, metronidazole, linezolid, fluconazole, voriconazole, azithromycin). To mimic real-time decision-making, clinicians (MD and PharmDs) retrospectively assessed IV to PO eligibility of the last IV antibiotic dose on admission for (1) GI/absorption, (2) clinical stability and (3) global review (but not given specific thresholds/criteria). Agreement level was compared for ID vs. non-ID reviews.

Results

Overall, 52 patients’ IV to PO eligibility was assessed by multiple clinicians; 5 GIM teams and 6 ID MDs or PharmDs participated. ID vs. GIM respective assessment of Global eligibility was 61% vs. 48% (agreement in 71% of cases). ID vs. GIM assessment of acceptable absorption was 82% vs. 67%; acceptable clinical stability was 64% vs. 62% (Fig 1). Clinician comments were reviewed to identify algorithm improvements and areas for frontline education.

Conclusion

Our results are consistent with prior data suggesting up to 40–50% of patients may be eligible for IV to PO conversion, even at institutions that have IV to PO protocols. Our data also shows that overall, ID clinicians were more likely to assess a patient as ready for PO antibiotics vs. non-ID clinicians. Our findings are important as understanding cases of non-agreement and obtaining GIM consensus for tool utility are important for our next step, assessing INFORMER implementation on realtime IV to PO conversion rates.

Disclosures

All authors: No reported disclosures.

Behaviors Indicative of Pain in Brain-Injured Adult Patients With Different Levels of Consciousness in the Intensive Care Unit

CONTEXT

Many brain-injured patients are unable to self-report their pain during their hospitalization in the intensive care unit (ICU), and existing behavioral pain scales may not be well suited.

OBJECTIVES

The objectives of this study were to describe and compare behaviors in brain-injured patients with different levels of consciousness during nociceptive and nonnociceptive care procedures in the ICU and to examine interrater agreement of individual behaviors as well as discriminative and criterion validation of putative pain behaviors.

METHODS

Brain-injured ICU patients were observed using a 40-item behavioral checklist before and during soft touch (i.e., nonnociceptive procedure), turning, and other care procedures (nociceptive) by pairs of trained raters. When possible, patients self-reported their pain on a 0-10 visual thermometer. Patients were classified into unconscious (Glasgow Coma Scale, 3<GCS≤8), altered consciousness (9≤GCS≤12), or conscious (13≤GCS≤15).

RESULTS

A sample of 147 patients participated (65 conscious, 56 altered consciousness, and 26 unconscious). Active behaviors (e.g., face expressions and body movements) were more frequent in conscious patients. High-percentage interrater agreement (80%-98%) was obtained for most behaviors. The total number of active behaviors was significantly higher during turning and other nociceptive procedures compared with rest (Wilcoxon = 9.873, P < 0.001) and soft touch (Wilcoxon = 9.486, P < 0.001) regardless of levels of consciousness. The strongest predictors of pain intensity (n = 33) were grimace, mouth opening, orbit tightening, eye weeping, and eyes tightly closed; these behaviors were moderately correlated with self-reported pain intensity (Spearman rho = 0.47).

CONCLUSION

These findings may guide the revision of existing pain scales to make their content more suited for this population.

Pros and Cons of 19 Sport-Related Concussion Educational Resources in Canada: Avenues for Better Care and Prevention

Objective: The goal of this research was to assess the effectiveness of available concussion educational resources in Canada, the means used to disseminate this knowledge and the impact of these educational resources on players' concussion prevention knowledge. Methods: We assessed concussion knowledge before and after exposure to one or more of 19 resources introduced through a national program aimed to increase awareness and knowledge of concussion. The effectiveness of the mode of delivery was measured by changes in concussion knowledge scores (CKS) between pre and pro scores. Measures: Concussion knowledge scores (CKS) were calculated for pre- and post- exposure to concussion educational resources and used as a measure of both, the effectiveness of each resource as well as the effectiveness of the delivery method. The effectiveness of each educational resource was also measured by the respondents' rating of each concussion educational resource. Results: Respondents in post-survey had higher CKS than those in pre-survey. Two out of the 19 newly developed concussion educational resources were effective in improving the resource users' CKS. Linear regression showed that using more resources further increased CKS. Four out of six modes of delivery enhanced respondents' concussion knowledge. Conclusion: Our findings demonstrate that the newly developed Canadian concussion educational resources were effective at improving users' concussion knowledge. Our data demonstrates that using three or more resources further enhanced the users' concussion knowledge. Future research, however, is critical to assess whether concussion prevention knowledge is sufficient to reduce injuries and factors influencing it.

Twitter and traumatic brain injury: A content and sentiment analysis of tweets pertaining to sport-related brain injury

Objectives:

Sport-related traumatic brain injuries are a significant public health burden, with hundreds of thousands sustained annually in North America. While sports offer numerous physical and social health benefits, traumatic brain injuries such as concussion can seriously impact a player’s life, athletic career, and sport enjoyment. The culture in many sports encourages winning at all costs, placing athletes at risk for traumatic brain injuries. As social media has become a central part of everyday life, the content of users’ messages often reflects the prevailing culture related to a particular event or health issue.

Methods:

We hypothesized that Twitter data might be useful for understanding public perceptions and misperceptions of sport-related traumatic brain injuries. We performed a content and sentiment analysis of 7483 Twitter^® tweets related to traumatic brain injuries in sports collected during June and July 2013.

Results:

We identified five major themes. Users tweeted about personal traumatic brain injuries experiences, reported traumatic brain injuries in professional athletes, shared research about sport-related concussions, and discussed policy and safety in injury prevention, such as helmet use. We identified mixed perceptions of and sentiment toward traumatic brain injuries in sports: both an understanding that brain injuries are serious and disregard for activities that might reduce the public burden of traumatic brain injuries were prevalent in our Twitter analysis.

Conclusion:

While the scientific and medical community considers a concussion a form of traumatic brain injuries, our study demonstrates a misunderstanding of this fact among the public. In our current digital age, social media can provide useful insight into the culture around a health issue, facilitating implementation of prevention and treatment strategies.

Factors affecting the concussion knowledge of athletes, parents, coaches, and medical professionals

OBJECTIVES

To determine the predictors of knowledge and awareness of concussion symptoms and outcomes through a survey of athletes, parents of players and coaches in sports settings in Canada.

METHODS

A cross-sectional survey of athletic communities in Canada was conducted. Respondents' concussion knowledge score consists of responses to questions about the symptoms, diagnosis, and treatment of a concussion and the timing of return-to-sport post-concussion. The percentage of correct responses was defined as the "identification rate." The extent to which participant factors affected the scores was examined by univariate and multivariate analyses.

RESULTS

Respondents were able to identify a mean of 80.6% of symptoms. Cognitive symptoms were most commonly known, and mental health symptoms associated with concussion were least known, and health professionals, coaches, and those with a personal history of concussion had the highest levels of overall knowledge. Language, age, educational level, annual household income, and traumatic brain injury history were good predictors of better concussion knowledge.

CONCLUSION

Those designing and implementing interventions aimed at concussion management and prevention should ensure that younger, lower income, lower educational, non-English-speaking persons, and those without experience of traumatic brain injury or concussion be specifically accounted for in the design and implementation of interventions to prevent and treat concussion and mild traumatic brain injury.

The Use of Social Media in Recruitment for Medical Research Studies: A Scoping Review

Background

Recruiting an adequate number of participants into medical research studies is challenging for many researchers. Over the past 10 years, the use of social media websites has increased in the general population. Consequently, social media websites are a new, powerful method for recruiting participants into such studies.

Objective

The objective was to answer the following questions: (1) Is the use of social media more effective at research participant recruitment than traditional methods? (2) Does social media recruit a sample of research participants comparable to that recruited via other methods? (3) Is social media more cost-effective at research participant recruitment than traditional methods?

Methods

Using the MEDLINE, PsycINFO, and EMBASE databases, all medical research studies that used social media and at least one other method for recruitment were identified. These studies were then categorized as either interventional studies or observational studies. For each study, the effectiveness of recruitment, demographic characteristics of the participants, and cost-effectiveness of recruitment using social media were evaluated and compared with that of the other methods used. The social media sites used in recruitment were identified, and if a study stated that the target population was “difficult to reach” as identified by the authors of the study, this was noted.

Results

Out of 30 studies, 12 found social media to be the most effective recruitment method, 15 did not, and 3 found social media to be equally effective as another recruitment method. Of the 12 studies that found social media to be the best recruitment method, 8 were observational studies while 4 were interventional studies. Of the 15 studies that did not find social media to be the best recruitment method, 7 were interventional studies while 8 were observational studies. In total, 8 studies stated that the target population was “hard-to-reach,” and 6 of these studies found social media to be the most effective recruitment method. Out of 14 studies that reported demographic data for participants, 2 studies found that social media recruited a sample comparable to that recruited via traditional methods and 12 did not. Out of 13 studies that reported cost-effectiveness, 5 studies found social media to be the most cost-effective recruitment method, 7 did not, and 1 study found social media equally cost-effective as compared with other methods.

Conclusions

Only 12 studies out of 30 found social media to be the most effective recruitment method. There is evidence that social media can be the best recruitment method for hard-to-reach populations and observational studies. With only 30 studies having compared recruitment through social media with other methods, more studies need to be done that report the effectiveness of recruitment for each strategy, demographics of participants recruited, and cost-effectiveness of each method.

Aggression, Violence and Injury in Minor League Ice Hockey: Avenues for Prevention of Injury

Background

In North America, more than 800,000 youth are registered in organized ice hockey leagues. Despite the many benefits of involvement, young players are at significant risk for injury. Body-checking and aggressive play are associated with high frequency of game-related injury including concussion. We conducted a qualitative study to understand why youth ice hockey players engage in aggressive, injury-prone behaviours on the ice.

Methods

Semi-structured interviews were conducted with 61 minor ice hockey participants, including male and female players, parents, coaches, trainers, managers and a game official. Players were aged 13–15 playing on competitive body checking teams or on non-body checking teams. Interviews were manually transcribed, coded and analyzed for themes relating to aggressive play in minor ice hockey.

Results

Parents, coaches, teammates and the media exert a large influence on player behavior. Aggressive behavior is often reinforced by the player’s social environment and justified by players to demonstrate loyalty to teammates and especially injured teammates by seeking revenge particularly in competitive, body-checking leagues. Among female and male players in non-body checking organizations, aggressive play is not reinforced by the social environment. These findings are discussed within the framework of social identity theory and social learning theory, in order to understand players’ need to seek revenge and how the social environment reinforces aggressive behaviors.

Conclusion

This study provides a better understanding of the players’ motivations and environmental influences around aggressive and violent play which may be conducive to injury. The findings can be used to help design interventions aimed at reducing aggression and related injuries sustained during ice hockey and sports with similar cultures and rules.

Inflammatory cytokine and chemokine profiles are associated with patient outcome and the hyperadrenergic state following acute brain injury

Background

Traumatic brain injury (TBI) elicits intense sympathetic nervous system (SNS) activation with profuse catecholamine secretion. The resultant hyperadrenergic state is linked to immunomodulation both within the brain and systemically. Dysregulated inflammation post-TBI exacerbates secondary brain injury and contributes to unfavorable patient outcomes including death. The aim of this study was to characterize the early dynamic profile of circulating inflammatory cytokines/chemokines in patients admitted for moderate-to-severe TBI, to examine interrelationships between these mediators and catecholamines, as well as clinical indices of injury severity and neurological outcome.

Methods

Blood was sampled from 166 isolated TBI patients (aged 45 ± 20.3 years; 74.7 % male) on admission, 6-, 12-, and 24-h post-injury and from healthy controls (N = 21). Plasma cytokine [interleukin (IL)-1β, -2, -4, -5, -10, -12p70, -13, tumor necrosis factor (TNF)-α, interferon (IFN)-γ] and chemokine [IL-8, eotaxin, eotaxin-3, IFN-γ-induced protein (IP)-10, monocyte chemoattractant protein (MCP)-1, -4, macrophage-derived chemokine (MDC), macrophage inflammatory protein (MIP)-1β, thymus activation regulated chemokine (TARC)] concentrations were analyzed using high-sensitivity electrochemiluminescence multiplex immunoassays. Plasma catecholamines [epinephrine (Epi), norepinephrine (NE)] were measured by immunoassay. Neurological outcome at 6 months was assessed using the extended Glasgow outcome scale (GOSE) dichotomized as good (>4) or poor (≤4) outcomes.

Results

Patients showed altered levels of IL-10 and all chemokines assayed relative to controls. Significant differences in a number of markers were evident between moderate and severe TBI cohorts. Elevated IL-8, IL-10, and TNF-α, as well as alterations in 8 of 9 chemokines, were associated with poor outcome at 6 months. Notably, a positive association was found between Epi and IL-1β, IL-10, Eotaxin, IL-8, and MCP-1. NE was positively associated with IL-1β, IL-10, TNF-α, eotaxin, IL-8, IP-10, and MCP-1.

Conclusions

Our results provide further evidence that exaggerated SNS activation acutely after isolated TBI in humans may contribute to harmful peripheral inflammatory cytokine/chemokine dysregulation. These findings are consistent with a potentially beneficial role for therapies aimed at modulating the inflammatory response and hyperadrenergic state acutely post-injury.

Electronic supplementary material

The online version of this article (doi:10.1186/s12974-016-0500-3) contains supplementary material, which is available to authorized users.

Antisaccadic Eye Movements Are Correlated with Corpus Callosum White Matter Mean Diffusivity, Stroop Performance, and Symptom Burden in Mild Traumatic Brain Injury and Concussion

Antisaccades are thought to involve higher level inputs from neural centers involved in rapid eye movement inhibition and control. Previous work has demonstrated that performance on the antisaccade task can help in the assessment of injury in acute and/or chronic mild traumatic brain injury (mTBI). In this exploratory study, we performed cross-sectional and longitudinal comparisons of rapid eye movement, followed by correlations of antisaccade performance with assessments of symptom burden, diffusion tensor imaging, and a neuropsychological test of response inhibition. Significant deficits in antisaccade median latency, F(2, 31) = 3.65, p = 0.04 and prosaccade error mean duration, F(2, 31) = 3.63, p = 0.04 were found between patient groups and controls: the former was correlated with loss of white matter integrity in the splenium of the corpus callosum in acute mTBI, rho = 0.90, p = 0.0005. Furthermore, increased antisaccade median latency was also correlated with poor performance on an executive functioning task, r² = 0.439, p = 0.03, and greater symptom burden, r² = 0.480, p = 0.02 in the acute mTBI patients. Our preliminary research suggests that the antisaccade task could be useful as a neurological marker for mTBI and concussion, but more work is required.

Web-based intervention programs for depression: a scoping review and evaluation

Background

Although depression is known to affect millions of people worldwide, individuals seeking aid from qualified health care professionals are faced with a number of barriers to treatment including a lack of treatment resources, limited number of qualified service providers, stigma associated with diagnosis and treatment, prolonged wait times, cost, and barriers to accessibility such as transportation and clinic locations. The delivery of depression interventions through the Internet may provide a practical solution to addressing some of these barriers.

Objective

The purpose of this scoping review was to answer the following questions: (1) What Web-delivered programs are currently available that offer an interactive treatment component for depression?, (2) What are the contents, accessibility, and usability of each identified program?, and (3) What tools, supports, and research evidence are available for each identified program?

Methods

Using the popular search engines Google, Yahoo, and Bing (Canadian platforms), two reviewers independently searched for interactive Web-based interventions targeting the treatment of depression. The Beacon website, an information portal for online health applications, was also consulted. For each identified program, accessibility, usability, tools, support, and research evidence were evaluated and programs were categorized as evidence-based versus non-evidence-based if they had been the subject of at least one randomized controlled trial. Programs were scored using a 28-point rating system, and evidence- versus non-evidence-based programs were compared and contrasted. Although this review included all programs meeting exclusion and inclusion criteria found using the described search method, only English language Web-delivered depression programs were awarded an evaluation score.

Results

The review identified 32 programs meeting inclusion criteria. There was a great deal of variability among the programs captured in this evaluation. Many of the programs were developed for general adolescent or adult audiences, with few (n=2) focusing on special populations (eg, military personnel, older adults). Cognitive behavioral therapy was the most common therapeutic approach used in the programs described. Program interactive components included mood assessments and supplementary homework sheets such as activity planning and goal setting. Only 12 of the programs had published evidence in support of their efficacy and treatment of depressive symptoms.

Conclusions

There are a number of interactive depression interventions available through the Internet. Recommendations for future programs, or the adaptation of existing programs include offering a greater selection of alternative languages, removing registration restrictions, free trial periods for programs requiring user fees, and amending programs to meet the needs of special populations (eg, those with cognitive and/or visual impairments). Furthermore, discussion of specific and relevant topics to the target audience while also enhancing overall user control would contribute to a more accessible intervention tool.

Memory impairment among people who are homeless: a systematic review

Cognitive impairment may interfere with an individual's ability to function independently in the community and may increase the risk of becoming and remaining homeless. The purpose of this study was to systematically review the literature on memory deficits among people who are homeless in order to gain a better understanding of its nature, causes and prevalence. Studies that measured memory functioning as an outcome among a sample of homeless persons were included. Data on sampling, outcome measures, facet of memory explored and prevalence of memory impairment were extracted from all selected research studies. Included studies were evaluated using a critical appraisal process targetted for reviewing prevalence studies. Eleven studies were included in the review. Verbal memory was the most commonly studied facet of memory. Potential contributing factors to memory deficits among persons who are homeless were explored in seven studies. Memory deficits were common among the samples of homeless persons studied. However, there was a great deal of variation in the methodology and quality of the included studies. Conceptualisations of "homelessness" also differed across studies. There is a need for more controlled research using validated neuropsychological tools to evaluate memory impairment among people who are homeless.

Traumatic brain injury among men in an urban homeless shelter: observational study of rates and mechanisms of injury

BACKGROUND

Little empiric research has investigated the interrelationship between homelessness and traumatic brain injury. The objectives of this study were to determine the rate, mechanisms and associated outcomes of traumatic brain injury among men in an urban homeless shelter.

METHODS

We recruited participants from an urban men's shelter in Toronto, Ontario. Researchers administered the Brain Injury Screening Questionnaire, a semistructured interview screening tool for brain injury. Demographic information and detailed histories of brain injuries were obtained. Participants with positive and negative screening results were compared, and the rates and mechanisms of injury were analyzed by age group.

RESULTS

A total of 111 men (mean age 54.2 ± standard deviation 11.5 yr; range 27-81 yr) participated. Nearly half (50 [45%]) of the respondents had a positive screening result for traumatic brain injury. Of these, 73% (35/48) reported experiencing their first injury before adulthood (< 18 yr), and 87% (40/46) reported a first injury before the onset of homelessness. Among those with a positive screening result, 33 (66%) reported sustaining at least one traumatic brain injury by assault, 22 (44%) by sports or another recreational activity, 21 (42%) by motor vehicle collision and 21 (42%) by a fall. A positive screening result was significantly associated with a lifetime history of arrest or mental illness and a parental history of substance abuse.

INTERPRETATION

Multiple mechanisms contributed to high rates of traumatic brain injury within a sample of homeless men. Assault was the most common mechanism, with sports and recreation, motor vehicle collisions and falls also being reported frequently by the participants. Injury commonly predated the onset of homelessness, with most participants experiencing their first injury in childhood. Additional research is needed to understand the complex interactions among homelessness, traumatic brain injury, mental illness and substance use.

Can fluctuations in vital signs be used for pain assessment in critically ill patients with a traumatic brain injury?

Background. Many critically ill patients with a traumatic brain injury (TBI) are unable to communicate. While observation of behaviors is recommended for pain assessment in nonverbal populations, they are undetectable in TBI patients who are under the effects of neuroblocking agents. Aim. This study aimed to validate the use of vital signs for pain detection in critically ill TBI patients. Methods. Using a repeated measure within subject design, participants (N = 45) were observed for 1 minute before (baseline), during, and 15 minutes after two procedures: noninvasive blood pressure: NIBP (nonnociceptive) and turning (nociceptive). At each assessment, vital signs (e.g., systolic, diastolic, mean arterial pressure (MAP), heart rate (HR), respiratory rate (RR), capillary saturation (SpO2), end-tidal CO2, and intracranial pressure (ICP)) were recorded. Results. Significant fluctuations (P < 0.05) in diastolic (F = 6.087), HR (F = 3.566), SpO2 (F = 5.740), and ICP (F = 3.776) were found across assessments, but they were similar during both procedures. In contrast, RR was found to increase exclusively during turning (t = 3.933; P < 0.001) and was correlated to participants' self-report. Conclusions. Findings from this study support previous ones that vital signs are not specific for pain detection. While RR could be a potential pain indicator in critical care, further research is warranted to support its validity in TBI patients with different LOC.

Traumatic brain injury among people who are homeless: a systematic review

Background

Homelessness and poverty are important social problems, and reducing the prevalence of homelessness and the incidence of injury and illness among people who are homeless would have significant financial, societal, and individual implications. Recent research has identified high rates of traumatic brain injury (TBI) among this population, but to date there has not been a review of the literature on this topic. The objective of this systematic review was to review the current state of the literature on TBI and homelessness in order to identify knowledge gaps and direct future research.

Methods

A systematic literature search was conducted in PsycINFO (1887–2012), Embase (1947–2012), and MEDLINE/Pubmed (1966–2012) to identify all published research studies on TBI and homelessness. Data on setting, sampling, outcome measures, and rate of TBI were extracted from these studies.

Results

Eight research studies were identified. The rate of TBI among samples of persons who were homeless varied across studies, ranging from 8%-53%. Across the studies there was generally little information to adequately describe the research setting, sample sizes were small and consisted mainly of adult males, demographic information was not well described, and validated screening tools were rarely used. The methodological quality of the studies included was generally moderate and there was little information to illustrate that the studies were adequately powered or that study samples were representative of the source population. There was also an absence of qualitative studies in the literature.

Conclusions

The rate of TBI is higher among persons who are homeless as compared to the general population. Both descriptive and interventional studies of individuals who are homeless should include a psychometrically sound measure of history of TBI and related disability. Education of caregivers of persons who are at risk of becoming, or are homeless, should involve training on TBI. Dissemination of knowledge to key stakeholders such as people who are homeless, their families, and public policy makers is also advocated.

Evaluation of an online cognitive behavioural therapy program by patients with traumatic brain injury and depression

PRIMARY OBJECTIVE

The most frequently reported psychiatric symptom after traumatic brain injury (TBI) is depression. This study examined whether internet-delivered cognitive behaviour therapy (CBT) could be appropriate and effective for patients with mild or moderate TBI and depression.

METHODS AND PROCEDURES

Patients were recruited for an at-home, 6-week internet-based CBT program (MoodGYM). Participants were assessed during this period by weekly telephone calls and at 12 months post-enrolment. Intervention completion rates, predictors of adherence, user feedback and changes in scores on validated depression scales were assessed.

MAIN OUTCOMES AND RESULTS

Twenty-one patients were recruited: 64% and 43% completed the 6-week intervention and the 12-month follow-up, respectively. Adherence rates were not predicted by demographic or injury characteristics in this small sample. Patients identified reading, memory and comprehension requirements as limitations of the program. Scores on the depression scales were significantly decreased upon completion of the intervention and at the 12-month follow-up.

CONCLUSIONS

The MoodGYM program may be effective for treating symptoms of depression in patients with TBI. While adherence rates were not predicted by age, education level or injury severity, demands upon memory and concentration which may already be compromised in these patients need to be considered.

Patient satisfaction and documentation of pain assessments and management after implementing the adult nonverbal pain scale

BACKGROUND

Accurate assessment and management of pain in critically ill patients who are nonverbal or cognitively impaired is challenging. No widely accepted assessment tool is currently in place for assessing pain in these patients.

OBJECTIVES

To evaluate the effect of implementing a new pain assessment tool in a trauma/neurosurgery intensive care unit.

METHODS

Staff and patient satisfaction questionnaires and retrospective chart reviews were used before and after implementation of the Nonverbal Pain Scale. The questionnaire responses, frequency of pain documentation, and amount of pain medication given were compared from before to after implementation.

RESULTS

Most staff (78%) ranked the tool as easy to use. Implementation of the tool increased staff confidence in assessing pain in nonverbal, sedated patients (57% before vs 81% after implementation, P = .02) and increased the number of pain assessments documented by the nursing staff for noncommunicative patients per day in the intensive care unit (2.2 before vs 3.4 after, P = .02). Patients reported decreased retrospective pain ratings (8.5 before vs 7.2 after, P = .04) and a trend toward a decrease in the time required to receive pain medication (38% before vs 10% after requiring >5 minutes to receive medication, P = .06).

CONCLUSIONS

Implementation of the Nonverbal Pain Scale in a critical care setting improved patients' ratings of their pain experience, improved documentation by nurses, and increased nurses' confidence in assessing pain in nonverbal patients.